Fix NLTK data permissions and Docker user setup

Dockerfile

@@ -5,7 +5,8 @@ ENV DEBIAN_FRONTEND=noninteractive \
     PYTHONUNBUFFERED=1 \
     PYTHONDONTWRITEBYTECODE=1 \
     PORT=7860 \
-    BASE_DIR=/app
+    BASE_DIR=/app \
+    NLTK_DATA=/app/nltk_data
 
 # Install Python and other dependencies
 RUN apt-get update && apt-get install -y --no-install-recommends \
@@ -20,25 +21,32 @@ RUN apt-get update && apt-get install -y --no-install-recommends \
     curl \
     && rm -rf /var/lib/apt/lists/*
 
+# Create a non-root user
+RUN useradd -m -u 1000 user
+
 # Set working directory
 WORKDIR /app
 
-# Create necessary directories
-RUN mkdir -p /app/data/pdfs /app/data/texts /app/data/index
+# Create necessary directories and set permissions
+RUN mkdir -p /app/data/pdfs /app/data/texts /app/data/index /app/nltk_data && \
+    chown -R user:user /app
 
 # Copy requirements first for better caching
-COPY requirements.txt .
+COPY --chown=user:user requirements.txt .
+
+# Switch to non-root user
+USER user
 
 # Install Python packages
-RUN pip3 install --no-cache-dir -r requirements.txt
-RUN pip3 install --no-cache-dir gradio
+RUN pip3 install --no-cache-dir --user -r requirements.txt
+RUN pip3 install --no-cache-dir --user gradio
 
 # Copy your application code and data
-COPY app/ /app/
-COPY data/ /app/data/
+COPY --chown=user:user app/ /app/
+COPY --chown=user:user data/ /app/data/
 
-# Create a simple Gradio interface for the API
-COPY spaces_app.py /app/
+# Copy your Gradio interface
+COPY --chown=user:user spaces_app.py /app/
 
 # Expose the port Hugging Face Spaces expects
 EXPOSE 7860

app/main.py

@@ -414,29 +414,34 @@ logger = logging.getLogger(__name__)
 # Initialize global variables in app state
 @app.on_event("startup")
 async def startup_event():
-    try:
-        logger.info("Starting application initialization...")
-        
-        # Download NLTK data
-        logger.info("Downloading NLTK data...")
-        try:
-            await asyncio.to_thread(nltk.download, 'punkt', quiet=True)
-            await asyncio.to_thread(nltk.download, 'punkt_tab', quiet=True)
-        except Exception as e:
-            logger.error(f"Error downloading NLTK data: {str(e)}")
-            raise
-            
-        # Initialize the model
-        logger.info("Initializing model...")
-
-        # Check if CUDA is available
-        device = "cuda" if torch.cuda.is_available() else "cpu"
-        logger.info(f"Using device: {device}")
+     # Check if CUDA is available
+    device = "cuda" if torch.cuda.is_available() else "cpu"
+    logger.info(f"Using device: {device}")
 
-        if device == "cpu":
-            logger.warning("GPU not detected. Model will run slower on CPU.")
+    if device == "cpu":
+        logger.warning("GPU not detected. Model will run slower on CPU.")
+            
+    """Initialize the application on startup."""
+    logger = logging.getLogger(__name__)
+    logger.info("Starting application initialization...")
+    
+    # Set NLTK data path
+    nltk_data_dir = os.environ.get('NLTK_DATA', os.path.join(os.path.expanduser('~'), 'nltk_data'))
+    os.makedirs(nltk_data_dir, exist_ok=True)
+    
+    # Download NLTK data
+    logger.info("Downloading NLTK data...")
+    try:
+        await asyncio.to_thread(nltk.download, 'punkt', download_dir=nltk_data_dir, quiet=True)
+        await asyncio.to_thread(nltk.download, 'punkt_tab', download_dir=nltk_data_dir, quiet=True)
+    except Exception as e:
+        logger.error(f"Error downloading NLTK data: {str(e)}")
+        raise Exception(f"Failed to initialize application: {str(e)}")
 
-        app.state.pipe = pipeline(
+    # Initialize the model and index
+    try:
+        global model, embedding_model, faiss_index, documents
+        model = pipeline(
             "text-generation",
             model=MODEL_ID,
             trust_remote_code=True,
@@ -444,16 +449,14 @@ async def startup_event():
             device_map="auto",
             torch_dtype=torch.float16 if device == "cuda" else torch.float32
         )
-        logger.info("Model initialized successfully")
-        
-        # Load or create index and related components
-        logger.info("Loading FAISS index and documents...")
-        app.state.index, app.state.documents, app.state.embedding_model = await load_or_create_index()
-        logger.info("FAISS index and documents loaded successfully")
+        embedding_model = SentenceTransformer('paraphrase-multilingual-MiniLM-L12-v2')
         
+        # Load or create the FAISS index
+        faiss_index, documents, embedding_model = await load_or_create_index()
+        logger.info("Application initialization completed successfully")
     except Exception as e:
-        logger.error(f"Failed to initialize application: {str(e)}")
-        raise
+        logger.error(f"Error initializing application: {str(e)}")
+        raise Exception(f"Failed to initialize application: {str(e)}")
 
 @app.post("/generate")
 async def generate_content(request: ContentRequest):
@@ -485,9 +488,9 @@ async def generate_content(request: ContentRequest):
             
         response = generate_response_with_rag(
             request.topic,  # Use topic as the prompt
-                    app.state.index,
-                    app.state.embedding_model,
-                    app.state.documents,
+                    faiss_index,
+                    embedding_model,
+                    documents,
             settings
                 )